Geochemistry, Geophysics, Geosystems

A new view of ridge segmentation and near-axis volcanism at the East Pacific Rise, 8°–12°N, from EM300 multibeam bathymetry



[1] New, high-resolution bathymetry for the East Pacific Rise between 8°N and 12°N was collected over a 6 km wide swath centered on the ridge axis using the 30 kHz Simrad EM300 multibeam system. The coverage area corresponds latitudinally to the designated Ridge2000 Integrated Studies Site (ISS) for fast spreading ridges. The EM300 data, gridded at 30 m latitude by 50 m longitude, represent a greater than 4X improvement in horizontal resolution over previously available multibeam data and a 2X improvement in depth resolution. The new bathymetry was used to update the locations and hierarchy of ridge offsets for this area. Among the many applications for this data, it enables us to tabulate volcanoes half the size that could be previously detected. The distribution of near-axis volcanic cones >25 m high suggests that this population of small, near-axis cones results from low effusion rate eruptions of the ridge axis.

1. Introduction

[2] The East Pacific Rise (EPR) between 8°N and 12°N is an area of intense investigation into the process of seafloor spreading, and was recently designated the type area for fast spreading rates by the Ridge 2000 Program (Figure 1). The advent of multibeam mapping systems in the early 1980s provided the first continuous coverage maps of the rise crest throughout this area [Macdonald et al., 1984]. The most comprehensive map for this area has ∼300 m grid cell size from transit satellite navigated SeaMARC II phase bathymetry [Macdonald et al., 1992]. SeaBeam coverage of the axis gridded at 100 m is available but is based on transit satellite navigation [Tighe et al., 1988]. Poor navigation from transit satellite and dithered GPS limited the resolution of these gridded data sets. Since then, progress on multibeam mapping has been made, but the areas mapped have been much more limited in extent (Figure 1). Smaller areas mapped by SeaBeam with P-code GPS navigation have been gridded at 80 m [Cochran et al., 1999; Wilcock et al., 1993]. The Simrad EM300 multibeam system represents a fundamental advance over previous systems in that it uses 30 kHz sound and 1° × 1° beam width, rather than 12 kHz and 2° or 2.67° beams, to ensonify the seafloor. The main purpose of this paper is to present a digital map series from the first EM300 survey of the EPR collected during 2005 aboard the R/V Thompson.

Figure 1.

Overview of bathymetry from the study area, and its regional setting with respect to major ridge offsets (created by GeoMapApp, The shaded-relief areas of the map indicate where 300 m bathymetry grids from surface ship measurements are available; the remainder of the map is derived from satellite altimetry. The major, first- and second-order, ridge axis discontinuities are labeled. The solid line box shows the area of the EM300 sonar swath. The bold dashed lines outline the area where previous higher-resolution (80 m gridded) multibeam bathymetry is available.

[3] The EM300 bathymetry represents an important advance over the previously available maps and bathymetric data. It is both high-resolution and spatially extensive, covering the entire length of ridge within the EPR Ridge 2000 Integrated Studies Site. Perhaps the most significant is that it is well navigated from continuous, undithered, differential GPS that has only been available since 1991. The new higher-resolution EM300 multibeam data set allows us to reexamine with the benefit of bathymetric data the small axial discontinuities picked previously from side-scan acoustic backscatter images. The new EM300 maps also can be used to examine detailed structure on the crest of the EPR at a scale that bridges the gap between local, highly detailed near-bottom mapping and regional, lower-resolution surface vessel sonar mapping. Two highlights that we comment on in this work are the hierarchy of ridge discontinuities and the distribution of small (>25 m high) near-axis volcanic cones.

2. Data Acquisition

[4] The Simrad EM300 multibeam sonar system on the R/V Thompson collected a single swath of bathymetry along the ridge crest of the EPR from 8°N to 12°N in November 2005 during a transit leg between San Diego, USA and Puerto Ayora, Galápagos. The entire survey was navigated using the C-Nav™ GcGPS, a real-time differential correction with submeter accuracy. The low horizontal dilution of precision, the stable series of position fixes, and a very favorable comparison between the new EM300 data and the transponder-net navigated ABE bathymetry from the EPR (V. Ferrini, personal communication, 2006) all suggest that navigation was excellent throughout the survey. The data were transmitted from ship to shore using HiSeasNet [Berger et al., 2006] during the survey, and was shown in preliminary form at the fall 2005 American Geophysical Union meeting while the cruise was still underway.

[5] The EM300 system operates at 30 kHz. This frequency allows for 5 m vertical resolution and smaller beam footprints compared to the standard 10–15 kHz multibeam systems aboard most research vessels. The EM300 system is a highly adaptable system, and uses active beam steering for control of the beam spacing to help maintain an even swath width and full bottom coverage. The positions of the beams were corrected for the ship's pitch, roll, heave, and yaw by a POS M/V inertial reference system. The sea state was fairly low during the survey, and despite having a headwind that caused the heading to deviate consistently from the course made good by ∼5° to the east, little beam steering was necessary. The average pitch was bow-down 0.2° ± 0.5° with a maximum angle of 2.3°. The average roll was 1.3° ± 0.7° to port with a maximum roll angle of 4.7° during the survey.

[6] The EM300 installed on the R/V Thompson transmits 135 beams per ping over a 120° angular swath using a steered beamforming array. The swath was ∼6 km during the survey, yielding an expected across-track beam spacing of ∼45 m. The EM300 uses both amplitude and phase detection to measure the received sound, thus the working across-track resolution is somewhere between the beam spacing and beam footprint. During data collection, the EM300 was operated in an equal-area mode, where the beamformer projects beams of varying angle across the swath to create equal size sonar footprints on the seafloor. The ping rate is automatically adjusted by the sonar to the limit of the round-trip traveltime of sound in the water. The average ping rate during the survey was 7.0 ± 0.1 seconds s. The survey was run at 9–10 knots (4.6–5.1 ms−1) speed over ground to maximize the survey area within time constraints imposed by the transit. Thus a 30–35 m along-track ping spacing is expected.

[7] The depth recorded by any echosounder is strongly dependent on the sound velocity profile through the water column. An expendable bathythermograph (XBT) probe was used to obtain a sound velocity profile through the thermocline at the start of the survey. The sound velocity profile from this XBT was applied to all pings during the survey.

[8] The data were processed using the MB-system software [Caress and Chayes, 1996]. We found the bathymetric readings to be sufficiently noise-free so that no automatic filters were applied. Some manual editing of the individual pings reduced the noise near the edges of the swath. The majority of the swath remained >5.5 km wide after editing. A radial Gaussian basis function was applied to the points to search for the optimal grid spacing using the method described by Ferrini et al. [2007]. From this analysis, it was determined that density of data points supported gridding at 30 m (along-track, north-south) by 50 m (across-track, east-west). This corresponds to our estimates of along-track and across-track data resolution based on physical parameters. Higher resolution might have been obtained by setting the across-track width of the sonar swath to a maximum of 5 km. Gridding was done using the mbgrid module of MB-system in the weighted beam footprint mode, which assigns a depth value based on the fraction of the grid cell that the footprint of the beam covers. Interpolation was applied using a minimum curvature spline to fill gaps of 1 grid cell between two defined grid cells. An examination of the gridded data indicates that each gridcell contains an average of 1.25 data points. Grids were made for 12°N to the Clipperton Transform and from Clipperton south to the Siquieros Transform (Figure 2). These grids reside at the Ridge2000 data portal (

Figure 2.

Shaded-relief bathymetric maps derived from 30 m × 50 m gridded EM300 depth measurements. Depth is plotted with a histogram-equalized color palette optimized for each panel to emphasize local changes in terrain. Locations of third-order (arrows) and fourth-order ridge discontinuities (lines) from Table 1 are indicated. Red stars indicate the location of each very-near-axis volcanic cone meeting the criteria explained in the text.

Figure 2.


Figure 2.


3. Structure of the Ridge Crest

[9] Ridge discontinuities are particularly important because they divide the ridge into discrete units of crustal accretion. Between 8°N and 12°N, the EPR exhibits a full range of scales of ridge discontinuities, but smaller offsets north and south of the 9°–10°N segment are only roughly defined from transit-satellite navigated SeaMARC II side-scan data [see Macdonald et al., 1992, Table 2]. These “fine-scale” offsets are subdivided into third or fourth order on the basis of criteria given by Macdonald et al. [1991] and White et al. [2000]. Although the resolution of the SeaMARC II side-scan data is similar to our EM300 bathymetry, improvements in navigational accuracy and the 3-dimensional perspective provided by the EM300 data permit a refinement of the previous segment boundary picks in this area (Figure 2).

[10] Table 1 presents a comprehensive list of axial discontinuities within the Ridge 2000 Integrated Study Site, including revisions to segmentation order and adjustments to locations of segment boundary offsets based on EM300 data presented here, and high-resolution side-scan data presented previously [Haymon et al., 1991; White et al., 2002]. Third-order offsets typically are subtle features in the 12 kHz multibeam records, but are much more obvious from EM300 (Figure 3a). Thus our picks in Table 1 for third-order ridge discontinuities are unlikely to be changed by further mapping. Fourth-order offsets, however, usually are invisible on 12 kHz multibeam maps, and sometimes are subtle on the EM300 maps (Figure 3b). Hence future surveys with higher-resolution, well-navigated side-scan sonars are likely to better define the location and structure of fourth-order ridge discontinuities outside the 9°–10°N segment where such data already exists [Fornari et al., 2004; Haymon et al., 1991; White et al., 2002].

Figure 3.

EM300 bathymetry of (left) third-order and (right) fourth-order ridge discontinuities. Contour interval is 5 m with a bold index contour at 25 m intervals. The third-order offset, at 8°37′N, separates an unusual segment with twin ridges to the north from a more typical segment to the south. The fourth-order discontinuity at 9°52′N is a very small right-stepping offset. Table 1 contains descriptions of these and all other ridge discontinuities in the survey area.

Table 1. Location of Ridge Discontinuities on the EPR, 8°–12°N From EM300 Bathymetrya
Latitude, NOrderOffset, kmSenseCommentSee Also
8°24′1140LSiquierosFornari et al. [1989]
8°37′32Roverlapping ridges; overlap zone extends 8°36′–39′N G, M 
8°45.5′30.4Rtransition from wide to narrow ASCT, discontinuity zone extends 8°45′–46′N M 
9°00.5′31.5Lend of ASCT 
9°03′210ROSC G, MKent et al. [2000]; Sempere and Macdonald [1986]
9°12′30.3R Haymon et al. [1991]; White et al. [2002]
9°20′30.9Rrevised location from 9°17′N [Haymon et al., 1991]; overlapping axial pillow ridges zone 9°21′–9°19′N G, S.White et al. [2002]
9°26′4<0.1-MHaymon et al. [1991]
9°32.7′4<0.1R Haymon et al. [1991]
9°34.9′4<0.1Rbend in bathymetryHaymon et al. [1991]
9°37′30.5Roverlapping ASCTs; overlap zone extends 9°36′–38′N MSmith et al. [2001]; Haymon et al. [1991]
9°44′3? Or 40.2Rpreviously identified as third order [White et al., 2002], but similar in total offset to other fourth order MHaymon et al. [1991]
9°49′40.1R Haymon et al. [1991]
9°53′40.2RCCW rotation of ridge axis; could be 9°51.5′ RAD [Haymon et al., 1991]? G 
9° 58′3>0.2Loverlapping axial pillow ridges; overlap zone extends 9°55.7′–59′N M?White et al. [2002]
10° 05′4<0.2Loffset of small ASCT M 
10°15′185RClippertonGallo et al. [1986]; Kastens et al. [1986]
10°32′30.7RGat 10°31′ in Macdonald et al. [1992]
10°35.5′40.2Rpoorly defined axial trace for 10°35′–36′N; several small offsets through this zone 
10°46.5′40-CCW rotation of ridge axis 
10°56′40.2Roverlapping ASCTs; overlap zone extends 10°55.8′–56.3′N 
10°59′30.3Roffset in ASCT; seismic AMC reflector to present north, absent to south [Detrick et al., 1987] 
11°01.5′40.1LOffset in ASCT 
11°08′40.2RG; S 
11°18.5′30.3Rat 11°19′ in Langmuir et al. [1986] G, M 
11°20.5′40-CW rotation of ridge axis and minor offset in ASCT 
11°25′4<0.1Loffset in ASCT 
11°28.5′30.3Loffset in ASCT G; S 
11°32′N4<0.1Laxis shifts left from narrow to broad ASCT 
11°45′29ROSC G, M, SPerram and Macdonald [1990]

4. Development of Very-Near-Axis Volcanic Cones

[11] Existing SeaBeam and SeaMARC II data from the EPR show volcanoes abruptly appearing on the ridge flanks at ∼5 km from the axis, having already achieved the median volume for the entire ridge flank volcano data set [Alexander and Macdonald, 1996; White et al., 1998]. However, only volcanoes >40 meters high can be reliably identified in these maps. Evidence exists from near-bottom observations for small-scale volcanism <2 km off-axis [Perfit et al., 1994; Schouten et al., 2003]. We also know that small volcanic mounds form on the crest of the ridge near ridge discontinuities [White et al., 2000, 2002]. This raises the question, how far off-axis do seamounts begin to form?

[12] The EM300 data sets a new threshold for the detection of seamounts since we can productively contour the data at 5 m instead of 10 m as used in previous studies. Thus we can now identify volcanoes as small as ∼25 m high (Figure 4). Volcanic cones were picked manually from digital maps contoured at 5 m depth intervals. All local highs defined by at least 5 contours were digitized from these maps. Only local highs with an aspect ratio of <2:1 length to width were retained as volcanic edifices in order to reduce the possibility of misidentifying upthrown fault blocks as volcanic cones. Several features that appear to be volcanic cones were excluded from this tabulation because the bathymetric contours do not close around the edifice. Most of the excluded features are small cones perched on the side of the axial high (Figure 4). Volcanoes were picked up to and including the Clipperton ridge-transform intersection, but were not picked within the transform. A digital elevation model created from the EM300 grids was used to calculate the height and volume of each digitized cone in our data set.

Figure 4.

EM300 bathymetry showing an example of two very-near-axis cones detected by this study (bold red outlines) and a cone perched on the side of the axial high (white arrow) that was not counted in this study. The closed-contour structure to the north of the outlined volcanoes fails to meet the criteria for volcanic cone because its ratio of length:width is >2. This is also the area of the 2003 eruption of the EPR [McClain et al., 2004]. Contour interval is 5 m with bold index contours at 25 m intervals.

[13] Unlike the isolated off-axis volcanoes farther from the ridge axis or beyond the first set of abyssal hills, some of the cones identified in this study may be rootless cones or hornitos. These are features that grow when lava flow inflation or collapse of a lava tube cause a cone to form far from the primary eruptive vent. Features of similar size have been interpreted as rootless cones on the MAR [Smith and Cann, 1999]. Unfortunately, distinguishing primary vents from rootless cones formed by lava flows pouring off of the ridge axis requires on- or near-bottom geologic mapping.

[14] A total of 86 very-near-axis volcanic cones were found within the study area (Figure 2). The maximum height for all cones in the area is 110 m, the mean height is 39 ± 15 m, and the median height is 35 m. Only 12 cones are >50 m high, large enough to have been counted on the older SeaBeam maps. Of these 12 cones, three are at ∼8°25′N, and the rest are between 10°N and 11°N. Approximately 30% (25 cones) lie within 1 km of the axis, thus within the axial neovolcanic zone. With only a ∼6 km wide swath, there is no particular significance to the across-axis distribution. To examine the along-axis distribution of these cones we summed the total number and total volume of cones in 10′ (18.5 km) latitude bins. These data were normalized for the total area of the EM300 swath within each bin.

[15] An interesting pattern emerges in the along-axis distribution of very-near-axis cones compared to major ridge offsets (Figure 5). The cones increase in number and total volume near the major ridge discontinuities at the Siquieros and Clipperton transforms, and at the 9°03′N and 11°45′N overlapping spreading centers (OSCs). These small cones are much more common north of Clipperton where the ridge is very narrow and erupts lower Mg basalts [Langmuir et al., 1986]. This relationship contrasts with prior regional observations that seamounts are larger and more abundant where the ridge is shallowest and broadest [Scheirer and Macdonald, 1995] (Figure 5). Possibly these small, very-near-axis cones reflect a change in volcanic style at the ridge axis from eruption of predominately lobate lava flows to pillow lava mounds near the ends of segments [White et al., 2000, 2002]. The greater abundance of pillow mounds where the ridge cross-section is narrower suggests that eruptions are both producing more mounds on-axis and burying fewer as they are rafted off-axis. We speculate that the population of volcanoes within 3 km of the axis is an expression of ridge segmentation, and part of the early growth of off-axis seamount chains.

Figure 5.

Comparison of the very-near-axis volcanism with the seamount chains farther off-axis. The top panel shows the volume of large off-axis seamount chains in ridge-perpendicular corridors extending from the axis to 1 Ma crust [Scheirer and Macdonald, 1995]. The bottom panel shows the abundance (columns) and volume (area under the curve) of very-near-axis volcanoes cataloged in this study. Local increases in both abundance and volume of very-near-axis volcanoes occur near major ridge axis discontinuities as labeled on the graph, whereas the seamounts show the opposite trend.

5. Conclusions

[16] Bathymetric maps of the entire EPR ridge crest within the Ridge 2000 Program Integrated Study Site across a 6 km wide swath are presented from EM300 multibeam data collected in November 2005 (see auxiliary material Animation S1). These maps are derived from data gridded at 30 m latitude by 50 m longitude, which was determined to be the finest grid interval supported by the data. This represents a greater than 4X improvement in gridcell resolution over previously available data over the ridge crest.

[17] The new EM300 bathymetry may be utilized in many different ways, to study the structure of the ridge axis or to serve as a geographic base for studies in other disciplines. Two new data products derived from the EM300 bathymetry and presented here are a revised list of ridge discontinuities within the EPR Integrated Study Site (Table 1) and the identification of 86 new volcanic cones, found near ridge discontinuities, within 3 km of the ridge axis.


[18] We thank Bill Martin, Mike Realander, and Rob Hagg of the UW marine technician group and Phil Smith of the R/V Thompson for assistance in collecting this data. This work benefited from discussions with Dan Fornari and Vicki Ferrini about the EM300 and Ken Macdonald about the ridge segmentation. Margo Edwards, Rob Sohn, and Doug Wilson provided insightful reviews that significantly improved the manuscript. Data collection was supported by NSF grants OCE-0326148 and OCE-0324668 and NOAA award NA04OAR4600049.